[Metrics] move prompt_tokens_total report to main process

Author: liyonghua0910

fastdeploy/engine/common_engine.py

@@ -1339,6 +1339,9 @@ def _insert_zmq_task_to_scheduler(self):
 
                         request.metrics.scheduler_recv_req_time = time.time()
                         main_process_metrics.inc_value("requests_number")
+                        main_process_metrics.inc_value("prompt_tokens_total", request.prompt_token_ids_len)
+                        main_process_metrics.obs_value("request_prompt_tokens", request.prompt_token_ids_len)
+                        main_process_metrics.obs_value("request_params_max_tokens", request.sampling_params.max_tokens)
                         trace_carrier = data.get("trace_carrier")
                         if trace_carrier:
                             request_id = get_base_request_id(data["request_id"])

fastdeploy/entrypoints/engine_client.py

@@ -53,7 +53,6 @@
     log_request,
     log_request_error,
 )
-from fastdeploy.metrics.metrics import main_process_metrics
 from fastdeploy.platforms import current_platform
 from fastdeploy.trace.constants import LoggingEventName
 from fastdeploy.trace.trace_logger import print as trace_print
@@ -374,9 +373,7 @@ async def add_requests(self, task):
 
             if "messages" in task:
                 task["messages"] = None
-            main_process_metrics.obs_value("request_params_max_tokens", task["max_tokens"])
-            main_process_metrics.inc_value("prompt_tokens_total", input_ids_len)
-            main_process_metrics.obs_value("request_prompt_tokens", input_ids_len)
+
         except Exception as e:
             log_request_error(
                 message="request[{request_id}] add_requests error: {error}, {traceback}",